#216 Position Aircraft (in air) crashes the sim

[Anonymous]

Originally created by: g...@av8n.com
Originally owned by: g...@av8n.com

Any attempt to use the dialog
        "Location" --> "Position Aircraft (in air)"
causes FG to abort.

This bug has existed (in various guises) for several years.
It is generally considered mostly a problem with JSBSIM
(rather than with the dialog GUI).

Console messages include the following.

passed invalid index (0) to FGRouteMgr::jumpToIndex

Trim Failed

  Trim Results:
       Angle of Attack:    5.5  wdot:  3.22e+01 Tolerance: 1e-03  Failed
              Throttle:    0.5  udot: -3.28e-01 Tolerance: 1e-03  Failed
            Pitch Trim:      0  qdot: -1.57e-09 Tolerance: 1e-04  Passed
Failed to get Mach number for given Vc and altitude, Vc unchanged.
Please mail the set of initial conditions used to apeden@earthlink.net

ESC[31mTried to initialize a non-existent engine!ESC[0m

Aircraft propulsion element has problems in file /games/orig/bogus/flightgear/Aircraft/c172p/c172p.xml
Unknown exception in the main loop. Aborting...
=====================================================================

commit [r1b4a4c9f52808308a90225a079cfcb505f194d5f]
Date:   Sun Dec 12 21:30:39 2010 +0100

[Anonymous]

Originally posted by: gijsrooy

Duplicate of issue #204.

Please use the search option before you file bugs. Searching for "Aircraft propulsion element has problems in file" was sufficient in this case.

Mergedinto: 204
Status: Duplicate

[Anonymous]

Originally posted by: zakalawe@mac.com

I think this is actually the same as bug 222, which for unfortunately was not being tracked here - since bug 204 is a recent, specific failure that appeared when Erik did the most recent sync of FG with the JSBSim mainline. The error above is certainly an instance of bug 204, but my assumption is that when bug 204 is fixed, bug 222 will re-appear again.

[Anonymous]

Originally posted by: g...@av8n.com

1) At least two people think this is not a duplicate of issue #204.
1a) The use-case for exhibiting the bug is not the same.
1b) The resulting symptoms are in some ways the same but in other
ways not the same;  for example the trim failure reported here
in connection with Location (in air) cannot possibly occur in
connection with Location (on ground) [which is the subject of bug #204].
1c) More importantly, the chronology is wrong.  This bug has been
around for years, as is well documented by the flightgear-devel
archives, whereas bug #204 seems to be newer, as zakalawe pointed
out.  That is, until rather recently, the Location (on ground)
feature wasn't a big problem.

2) This is also not the same as bug #222.  As for the root-cause analysis
given in connection with bug #222, part of it applies to this bug but part
does not.

2a) We agree that JSBSIM behaves badly when initialized for the Nth time
(for all N>1).  We agree that this strongly suggests some abuse of static
data.

2b) On the other hand, the part of the analysis that refers to dt=0, and
the headline of bug #222 that refers to NaN cannot possibly be the whole
story, and cannot possibly explain the symptoms reported here, because
the trim algorithm does not depend on dt, and neither does engine selection.
Engine selection is an integer concept, and NaN is a floating-point concept.

2c) The weird menagerie of symptoms suggests that there is /corruption/ of
memory.  We do not yet know which came first, the chicken or the egg.  That
is, memory corruption might be the /cause/ of the other bad things, or it
might be the /result/ of the other bad things.

My bet is that the code is buggy in at least two places.

It may be that the best way to proceed would be an extensive redesign of
the JSBSIM initialization section, but we don't know that for sure.  It
may be that fixing one of these three bugs will fix all the others, but
we don't know that for sure.  What they all have in common is JSBSIM
re-initialization.

Mergedinto: -204
Status: New

[Anonymous]

Originally posted by: bcoco... (code.google.com)@gmail.com

This bug should also be fixed by http://gitorious.org/fg/flightgear/commit/591ef4ab0157f79d35486a5c7ac1e16147cda7af

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

Fixed (JSBSim reset NaN issue). Caused by heap corruption resulting from properties not properly being untied in JSBSim.

Status: Fixed

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

(No comment was entered for this change.)

Status: Temp

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

(No comment was entered for this change.)

Status: Fixed

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

(No comment was entered for this change.)

Labels: Milestone-2.4.0